R Notebook

This is an R Markdown Notebook. When you execute code within the notebook, the results appear beneath the code.

Try executing this chunk by clicking the Run button within the chunk or by placing your cursor inside it and pressing Ctrl+Shift+Enter.

library(xgboost)
library(Matrix)
library(mclust)

    __  ___________    __  _____________
   /  |/  / ____/ /   / / / / ___/_  __/
  / /|_/ / /   / /   / / / /\__ \ / /   
 / /  / / /___/ /___/ /_/ /___/ // /    
/_/  /_/\____/_____/\____//____//_/    version 5.4.9
Type 'citation("mclust")' for citing this R package in publications.

library(tidyverse)

Registered S3 methods overwritten by 'dbplyr':
  method         from
  print.tbl_lazy     
  print.tbl_sql      
─ Attaching packages ─────────────────────────────────── tidyverse 1.3.1 ─
✓ tibble  3.1.5     ✓ stringr 1.4.0
✓ readr   2.0.2     ✓ forcats 0.5.1
✓ purrr   0.3.4     
─ Conflicts ───────────────────────────────────── tidyverse_conflicts() ─
x dplyr::collapse()   masks IRanges::collapse()
x dplyr::combine()    masks Biobase::combine(), BiocGenerics::combine()
x dplyr::count()      masks matrixStats::count()
x dplyr::desc()       masks IRanges::desc()
x tidyr::expand()     masks Matrix::expand(), S4Vectors::expand()
x plotly::filter()    masks dplyr::filter(), stats::filter()
x dplyr::first()      masks S4Vectors::first()
x widgetTools::funs() masks dplyr::funs()
x dplyr::lag()        masks stats::lag()
x purrr::map()        masks mclust::map()
x tidyr::pack()       masks Matrix::pack()
x ggplot2::Position() masks BiocGenerics::Position(), base::Position()
x purrr::reduce()     masks GenomicRanges::reduce(), IRanges::reduce()
x plotly::rename()    masks dplyr::rename(), S4Vectors::rename()
x purrr::simplify()   masks clusterProfiler::simplify(), DelayedArray::simplify()
x plotly::slice()     masks dplyr::slice(), xgboost::slice(), IRanges::slice()
x tidyr::unpack()     masks Matrix::unpack()

数值化

ds2训练分类器

ds2_data <- get_data_table(ds2, highvar = F, type = "data")
ds2_label <- as.numeric(as.character(Idents(ds2)))

index <- c(1:dim(ds2_data)[2]) %>% sample(ceiling(0.3*dim(ds2_data)[2]), replace = F, prob = NULL)
colnames(ds2_data) <- NULL

ds2_train_data <- list(data = t(as(ds2_data[,-index],"dgCMatrix")), label = ds2_label[-index])
ds2_test_data <- list(data = t(as(ds2_data[,index],"dgCMatrix")), label = ds2_label[index])

ds2_train <- xgb.DMatrix(data = ds2_train_data$data,label = ds2_train_data$label)
ds2_test <- xgb.DMatrix(data = ds2_test_data$data,label = ds2_test_data$label)

watchlist <- list(train = ds2_train, eval = ds2_test)
xgb_param <- list(eta = 0.2, max_depth = 6, 
                  subsample = 0.6,  num_class = length(table(Idents(ds2))),
                  objective = "multi:softprob", eval_metric = 'mlogloss')

bst_model <- xgb.train(xgb_param, ds2_train, nrounds = 100, watchlist, verbose = 0)

eval_loss <- bst_model[["evaluation_log"]][["eval_mlogloss"]]
plot_ly(data.frame(eval_loss), x = c(1:100), y = eval_loss) %>% 
  add_trace(type = "scatter", mode = "markers+lines", 
            marker = list(color = "black", line = list(color = "#1E90FFC7", width = 1)),
            line = list(color = "#1E90FF80", width = 2)) %>% 
  layout(xaxis = list(title = "epoch"),yaxis = list(title = "eval_mlogloss"))

ds2 -> ds1

Idents(ds1) <- ds1$seurat_clusters
temp <- get_data_table(ds1, highvar = F, type = "data")
ds1_data <- matrix(data=0, nrow = length(rownames(ds2_data)), ncol = length(colnames(temp)), 
                   byrow = FALSE, dimnames = list(rownames(ds2_data),colnames(temp)))
for(i in intersect(rownames(ds2_data), rownames(temp))){
  ds1_data[i,] <- temp[i,]
}
rm(temp)
ds1_label <- as.numeric(as.character(Idents(ds1)))
colnames(ds1_data) <- NULL
ds1_test_data <- list(data = t(as(ds1_data,"dgCMatrix")), label = ds1_label)
ds1_test <- xgb.DMatrix(data = ds1_test_data$data,label = ds1_test_data$label)

#预测结果

predict_ds1_test <- predict(bst_model, newdata = ds1_test)

predict_prop_ds1 <- matrix(data=predict_ds1_test, nrow = length(levels(Idents(ds2))), 
                           ncol = ncol(ds1), byrow = FALSE, 
                           dimnames = list(levels(Idents(ds2)),colnames(ds1)))

## 得到分群结果
ds1_res <- apply(predict_prop_ds1,2,func,rownames(predict_prop_ds1))
Idents(ds1) <- factor(ds1_res,levels = c(0:4))
umapplot(ds1)

ds1$supclustering <- Idents(ds1) #保存监督聚类结果

数值化地投射回umap

embedding <- FetchData(object = ds1, vars = c("UMAP_1", "UMAP_2"))
embedding <- cbind(embedding, t(predict_prop_ds1))

ggobj <- ggplot() +
  geom_point(data = embedding[embedding$`0`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `0`), shape=16, size = 3, alpha=0.5) + 
  scale_color_gradient('0', low = "#FFFFFF00", high = "#6dc0a6") +
  new_scale("color") +
    geom_point(data = embedding[embedding$`1`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `1`), size = 3, alpha=0.5) + 
  scale_color_gradient('1', low = "#FFFFFF00", high = "#e2b398") +
   new_scale("color") +
    geom_point(data = embedding[embedding$`2`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `2`), size = 2, alpha=0.5) + 
  scale_color_gradient('2', low = "#FFFFFF00", high = "#e2a2ca") +
  new_scale("color") +
    geom_point(data = embedding[embedding$`3`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `3`), size = 3, alpha=0.5) + 
  scale_color_gradient('3', low = "#FFFFFF00", high = "#d1eba8") +
   new_scale("color") +
      geom_point(data = embedding[embedding$`4`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `4`), size = 3, alpha=0.5) + 
  scale_color_gradient('4', low = "#FFFFFF00", high = "#b1d6fb") +
    new_scale("color") +
        xlab("UMAP 1") + ylab("UMAP 2")  +
        theme(axis.line = element_line(arrow = arrow(length = unit(0.2, "cm")))) +
        scale_y_continuous(breaks = NULL) +
        scale_x_continuous(breaks = NULL) + 
  theme(panel.background = element_blank(), panel.grid = element_blank(), legend.position = "bottom")
ggsave("pre_ds1_umap.svg",device = svg,plot = ggobj,height = 10,width = 10)

#ds2 -> ds0

embedding <- FetchData(object = ds0, vars = c("UMAP_1", "UMAP_2"))
embedding <- cbind(embedding, t(predict_prop_ds0))

ggobj <- ggplot() +
  geom_point(data = embedding[embedding$`0`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `0`), shape=16, size = 3, alpha=0.5) + 
  scale_color_gradient('0', low = "#FFFFFF00", high = "#6dc0a6") +
  new_scale("color") +
    geom_point(data = embedding[embedding$`1`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `1`),shape=16, size = 3, alpha=0.5) + 
  scale_color_gradient('1', low = "#FFFFFF00", high = "#e2b398") +
   new_scale("color") +
    geom_point(data = embedding[embedding$`2`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `2`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('2', low = "#FFFFFF00", high = "#e2a2ca") +
  new_scale("color") +
    geom_point(data = embedding[embedding$`3`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `3`),shape=16, size = 3, alpha=0.5) + 
  scale_color_gradient('3', low = "#FFFFFF00", high = "#d1eba8") +
   new_scale("color") +
      geom_point(data = embedding[embedding$`4`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `4`),shape=16, size = 3, alpha=0.5) + 
  scale_color_gradient('4', low = "#FFFFFF00", high = "#b1d6fb") +
    new_scale("color") +
        xlab("UMAP 1") + ylab("UMAP 2")  +
        theme(axis.line = element_line(arrow = arrow(length = unit(0.2, "cm")))) +
        scale_y_continuous(breaks = NULL) +
        scale_x_continuous(breaks = NULL) + 
  theme(panel.background = element_blank(), panel.grid = element_blank(), legend.position = "bottom")
ggsave("pre_ds1_umap.svg",device = svg,plot = ggobj,height = 10,width = 10)

PA -> AC

Idents(ds2_PA) <- ds2_PA$seurat_clusters
selected_features <- read.csv("./datatable/selected_features.csv", stringsAsFactors = F)
selected_features <- selected_features$x
PA_data <- get_data_table(ds2_PA, highvar = F, type = "data")
PA_data <- PA_data[selected_features,]
PA_label <- as.numeric(as.character(Idents(ds2_PA)))
colnames(PA_data) <- NULL

PA_train_data <- list(data = t(as(PA_data,"dgCMatrix")), label = PA_label)
PA_train <- xgb.DMatrix(data = PA_train_data$data,label = PA_train_data$label)
xgb_param <- list(eta = 0.2, max_depth = 6, 
                  subsample = 0.6,  num_class = length(table(Idents(ds2_PA))),
                  objective = "multi:softprob", eval_metric = 'mlogloss')

bst_model <- xgb.train(xgb_param, PA_train, nrounds = 100, verbose = 0)

embedding <- FetchData(object = ds2_AC, vars = c("UMAP_1", "UMAP_2"))
embedding <- cbind(embedding, t(predict_prop_AC))

ggobj <- ggplot() +
  geom_point(data = embedding[embedding$`0`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `0`), shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('0', low = "#FFFFFF00", high = "#6dc0a6") +
  new_scale("color") +
    geom_point(data = embedding[embedding$`1`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `1`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('1', low = "#FFFFFF00", high = "#e2b398") +
   new_scale("color") +
    geom_point(data = embedding[embedding$`2`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `2`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('2', low = "#FFFFFF00", high = "#e2a2ca") +
        xlab("UMAP 1") + ylab("UMAP 2")  +
        theme(axis.line = element_line(arrow = arrow(length = unit(0.2, "cm")))) +
        scale_y_continuous(breaks = NULL) +
        scale_x_continuous(breaks = NULL) + 
  theme(panel.background = element_blank(), panel.grid = element_blank(), legend.position = "bottom")
ggsave("ds2_PAtoAC_umap.svg",device = svg,plot = ggobj,height = 8,width = 8)

AC to PA

Idents(ds2_AC) <- ds2_AC$seurat_clusters
selected_features <- read.csv("./datatable/selected_features.csv", stringsAsFactors = F)
selected_features <- selected_features$x
AC_data <- get_data_table(ds2_AC, highvar = F, type = "data")
AC_data <- AC_data[selected_features,]
AC_label <- as.numeric(as.character(Idents(ds2_AC)))
colnames(AC_data) <- NULL

AC_train_data <- list(data = t(as(AC_data,"dgCMatrix")), label = AC_label)
AC_train <- xgb.DMatrix(data = AC_train_data$data,label = AC_train_data$label)
xgb_ACram <- list(eta = 0.2, max_depth = 6, 
                  subsample = 0.6,  num_class = length(table(Idents(ds2_AC))),
                  objective = "multi:softprob", eval_metric = 'mlogloss')

bst_model <- xgb.train(xgb_ACram, AC_train, nrounds = 100, verbose = 0)

Idents(ds2_PA) <- factor(ds2_PA$seurat_clusters,levels = c(0,1,2))

PA_data <- get_data_table(ds2_PA, highvar = F, type = "data")
PA_data <- PA_data[selected_features,]
PA_label <- as.numeric(as.character(Idents(ds2_PA)))
colnames(PA_data) <- NULL
PA_test_data <- list(data = t(as(PA_data,"dgCMatrix")), label = PA_label)
PA_test <- xgb.DMatrix(data = PA_test_data$data,label = PA_test_data$label)

#预测结果
predict_prop_PA <-predict(bst_model, newdata = PA_test) %>%
 matrix(nrow = length(levels(Idents(ds2_AC))), 
                           ncol = ncol(ds2_PA), byrow = FALSE, 
                           dimnames = list(levels(Idents(ds2_AC)),colnames(ds2_PA)))
PA_res <- apply(predict_prop_PA,2,func,rownames(predict_prop_PA))

confuse_matrix1 <- table(PA_test_data$label, PA_res, dnn=c("true","pre"))
sankey_plot(confuse_matrix1,session = "ACtoPA")

Idents(ds2_PA) <- factor(PA_res)
umapplot(ds2_PA)

embedding <- FetchData(object = ds2_PA, vars = c("UMAP_1", "UMAP_2"))
embedding <- cbind(embedding, t(predict_prop_PA))

ggobj <- ggplot() +
  geom_point(data = embedding[embedding$`0`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `0`), shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('0', low = "#FFFFFF00", high = "#6dc0a6") +
  new_scale("color") +
    geom_point(data = embedding[embedding$`1`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `1`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('1', low = "#FFFFFF00", high = "#e2b398") +
   new_scale("color") +
    geom_point(data = embedding[embedding$`2`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `2`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('2', low = "#FFFFFF00", high = "#e2a2ca") +
     new_scale("color") +
    geom_point(data = embedding[embedding$`3`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `3`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('3', low = "#FFFFFF00", high = "#d1eba8") +
        xlab("UMAP 1") + ylab("UMAP 2")  +
        theme(axis.line = element_line(arrow = arrow(length = unit(0.2, "cm")))) +
        scale_y_continuous(breaks = NULL) +
        scale_x_continuous(breaks = NULL) + 
  theme(panel.background = element_blank(), panel.grid = element_blank(), legend.position = "bottom")
ggsave("ds2_ACtoPA_umap.svg",device = svg,plot = ggobj,height = 8,width = 8)

在ds0上训练

Idents(ds0) <- ds0$seurat_clusters
ds0_data <- get_data_table(ds0, highvar = F, type = "data")
ds0_label <- as.numeric(as.character(Idents(ds0)))

index <- c(1:dim(ds0_data)[2]) %>% sample(ceiling(0.3*dim(ds0_data)[2]), replace = F, prob = NULL)
colnames(ds0_data) <- NULL

ds0_train_data <- list(data = t(as(ds0_data[,-index],"dgCMatrix")), label = ds0_label[-index])
ds0_test_data <- list(data = t(as(ds0_data[,index],"dgCMatrix")), label = ds0_label[index])

ds0_train <- xgb.DMatrix(data = ds0_train_data$data,label = ds0_train_data$label)
ds0_test <- xgb.DMatrix(data = ds0_test_data$data,label = ds0_test_data$label)

watchlist <- list(train = ds0_train, eval = ds0_test)
xgb_param <- list(eta = 0.2, max_depth = 6, 
                  subsample = 0.6,  num_class = length(table(Idents(ds0))),
                  objective = "multi:softprob", eval_metric = 'mlogloss')

bst_model <- xgb.train(xgb_param, ds0_train, nrounds = 100, watchlist, verbose = 0)

eval_loss <- bst_model[["evaluation_log"]][["eval_mlogloss"]]
plot_ly(data.frame(eval_loss), x = c(1:100), y = eval_loss) %>% 
  add_trace(type = "scatter", mode = "markers+lines", 
            marker = list(color = "black", line = list(color = "#1E90FFC7", width = 1)),
            line = list(color = "#1E90FF80", width = 2)) %>% 
  layout(xaxis = list(title = "epoch"),yaxis = list(title = "eval_mlogloss"))

importance <- xgb.importance(colnames(ds0_train), model = bst_model)
head(importance)

xgb.ggplot.importance(head(importance,20),n_clusters = 1) + theme_bw()+theme(
    axis.title.x = element_text(size = 15), axis.text.x = element_text(size = 8, colour = "black"),
    axis.title.y = element_text(size = 15), axis.text.y = element_text(size = 12, colour = "black"),
    legend.text = element_text(size = 20), legend.title = element_blank(), panel.grid = element_blank())

write.csv(importance, "./datatable/ds0_features.csv", row.names = F)
multi_featureplot(head(importance,9)$Feature, ds0, labels = "") 

Warning: Using `as.character()` on a quosure is deprecated as of rlang 0.3.0.
Please use `as_label()` or `as_name()` instead.
This warning is displayed once per session.

ds0 -> ds2

Idents(ds2) <- ds2$seurat_clusters 
temp <- get_data_table(ds2, highvar = F, type = "data")
ds2_data <- matrix(data=0, nrow = length(rownames(ds0_data)), ncol = length(colnames(temp)), 
                   byrow = FALSE, dimnames = list(rownames(ds0_data),colnames(temp)))
for(i in intersect(rownames(ds2_data), rownames(temp))){
  ds2_data[i,] <- temp[i,]
}
rm(temp)
ds2_label <- as.numeric(as.character(Idents(ds2)))
colnames(ds2_data) <- NULL
ds2_test_data <- list(data = t(as(ds2_data,"dgCMatrix")), label = ds2_label)
ds2_test <- xgb.DMatrix(data = ds2_test_data$data,label = ds2_test_data$label)

#预测结果

predict_ds2_test <- predict(bst_model, newdata = ds2_test)

predict_prop_ds2 <- matrix(data=predict_ds2_test, nrow = bst_model[["params"]][["num_class"]], 
                           ncol = ncol(ds2), byrow = FALSE, 
                           dimnames = list(c(0:(bst_model[["params"]][["num_class"]]-1)),colnames(ds2)))

## 得到分群结果
ds2_res <- apply(predict_prop_ds2,2,func,rownames(predict_prop_ds2))
confuse_matrix1 <- table(ds2_test_data$label, ds2_res, dnn=c("true","pre"))

sankey_plot(confuse_matrix1,0:5,0:4,session = "ds0tods2")

Idents(ds2) <- factor(ds2_res,levels = c(0:5))
umapplot(ds2)

embedding <- FetchData(object = ds2, vars = c("UMAP_1", "UMAP_2"))
embedding <- cbind(embedding, t(predict_prop_ds2))

ggobj <- ggplot() +
  geom_point(data = embedding[embedding$`0`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `0`), shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('0', low = "#FFFFFF00", high = "#6dc0a6") +
  new_scale("color") +
    geom_point(data = embedding[embedding$`1`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `1`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('1', low = "#FFFFFF00", high = "#e2b398") +
   new_scale("color") +
    geom_point(data = embedding[embedding$`2`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `2`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('2', low = "#FFFFFF00", high = "#e2a2ca") +
     new_scale("color") +
    geom_point(data = embedding[embedding$`3`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `3`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('3', low = "#FFFFFF00", high = "#d1eba8") +
     new_scale("color") +
    geom_point(data = embedding[embedding$`4`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `4`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('4', low = "#FFFFFF00", high = "#b1d6fb") +
     new_scale("color") +
    geom_point(data = embedding[embedding$`5`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `5`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('5', low = "#FFFFFF00", high = "#fd9999") +
        xlab("UMAP 1") + ylab("UMAP 2")  +
        theme(axis.line = element_line(arrow = arrow(length = unit(0.2, "cm")))) +
        scale_y_continuous(breaks = NULL) +
        scale_x_continuous(breaks = NULL) + 
  theme(panel.background = element_blank(), panel.grid = element_blank(), legend.position = "bottom")
ggsave("ds0tods2umap.svg",device = svg,plot = ggobj,height = 8,width = 8)

ds0 -> ds1

embedding <- FetchData(object = ds1, vars = c("UMAP_1", "UMAP_2"))
embedding <- cbind(embedding, t(predict_prop_ds1))

ggobj <- ggplot() +
  geom_point(data = embedding[embedding$`0`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `0`), shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('0', low = "#FFFFFF00", high = "#6dc0a6") +
  new_scale("color") +
    geom_point(data = embedding[embedding$`1`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `1`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('1', low = "#FFFFFF00", high = "#e2b398") +
   new_scale("color") +
    geom_point(data = embedding[embedding$`2`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `2`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('2', low = "#FFFFFF00", high = "#e2a2ca") +
     new_scale("color") +
    geom_point(data = embedding[embedding$`3`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `3`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('3', low = "#FFFFFF00", high = "#d1eba8") +
     new_scale("color") +
    geom_point(data = embedding[embedding$`4`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `4`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('4', low = "#FFFFFF00", high = "#b1d6fb") +
     new_scale("color") +
    geom_point(data = embedding[embedding$`5`>0.1,], 
             aes(x = UMAP_1, y = UMAP_2, color = `5`),shape=16, size = 2, alpha=0.5) + 
  scale_color_gradient('5', low = "#FFFFFF00", high = "#fd9999") +
        xlab("UMAP 1") + ylab("UMAP 2")  +
        theme(axis.line = element_line(arrow = arrow(length = unit(0.2, "cm")))) +
        scale_y_continuous(breaks = NULL) +
        scale_x_continuous(breaks = NULL) + 
  theme(panel.background = element_blank(), panel.grid = element_blank(), legend.position = "bottom")
ggsave("ds0tods1umap.svg",device = svg,plot = ggobj,height = 8,width = 8)

##lym

ARI 和聚类数的关系

Add a new chunk by clicking the Insert Chunk button on the toolbar or by pressing Ctrl+Alt+I.

When you save the notebook, an HTML file containing the code and output will be saved alongside it (click the Preview button or press Ctrl+Shift+K to preview the HTML file).

The preview shows you a rendered HTML copy of the contents of the editor. Consequently, unlike Knit, Preview does not run any R code chunks. Instead, the output of the chunk when it was last run in the editor is displayed.